PDF: Manipulate the selected text buffer to deal with special 0xfffe hyphens.

pdf/pdfium/pdfium_engine.cc

 // Copyright (c) 2012 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "pdf/pdfium/pdfium_engine.h"
 
 #include <math.h>
 
 #include "base/json/json_writer.h"
 #include "base/logging.h"
@@ skipping 27 matching lines @@
 #include "third_party/pdfium/fpdfsdk/include/fpdfedit.h"
 #include "third_party/pdfium/fpdfsdk/include/fpdfoom.h"
 #include "third_party/pdfium/fpdfsdk/include/fpdfppo.h"
 #include "third_party/pdfium/fpdfsdk/include/fpdfsave.h"
 #include "third_party/pdfium/fpdfsdk/include/pdfwindow/PDFWindow.h"
 #include "third_party/pdfium/fpdfsdk/include/pdfwindow/PWL_FontMap.h"
 #include "ui/events/keycodes/keyboard_codes.h"
 
 namespace chrome_pdf {
 
+namespace {
+
 #define kPageShadowTop    3
 #define kPageShadowBottom 7
 #define kPageShadowLeft   5
 #define kPageShadowRight  5
 
 #define kPageSeparatorThickness 4
 #define kHighlightColorR 153
 #define kHighlightColorG 193
 #define kHighlightColorB 218
 
@@ skipping 393 matching lines @@
     case 3:
       *offset_x = page_height - source_clip_box.right;
       *offset_y = page_width - source_clip_box.top;
       break;
     default:
       NOTREACHED();
       break;
   }
 }
 
-// Do an in-place transformation of objects on |page|. Translate all objects on
-// |page| in |source_clip_box| by (|offset_x|, |offset_y|) and scale them by
-// |scale_factor|.
-//
-// |page| Handle to the page. Returned by FPDF_LoadPage function.
-// |source_clip_box| specifies the source clip box positions, relative to
-// origin at left-bottom.
-// |scale_factor| specifies the scale factor that should be applied to page
-// objects.
-// |offset_x| and |offset_y| specifies the translation offsets for the page
-// objects, relative to origin at left-bottom.
+// This formats a string with special 0xfffe end-of-line hyphens the same way
+// as Adobe Reader. When a hyphen is encountered, the next non-CR/LF whitespace
+// becomes CR+LF and the hyphen is erased. If there is no whitespace between
+// two hyphens, the latter hyphen is erased and ignored.
+void FormatStringWithHyphens(base::string16* text) {
+  // First pass marks all the hyphen positions.
+  struct HyphenPosition {
+    HyphenPosition() : position(0), next_whitespace_position(0) {}
+    size_t position;
+    size_t next_whitespace_position;  // 0 for none
+  };
+  std::vector<HyphenPosition> hyphen_positions;
+  HyphenPosition current_hyphen_position;
+  bool current_hyphen_position_is_valid = false;
+  const base::char16 kPdfiumHyphenEOL = 0xfffe;
 
-void TransformPageObjects(FPDF_PAGE page, const ClipBox& source_clip_box,

[Lei Zhang, 2014/09/11 23:13:10]

This has been unused for some time. Putting this big block in an anonymous
namespace made clang warn me about it.

-                          const double scale_factor, double offset_x,
-                          double offset_y) {
-  const int obj_count = FPDFPage_CountObject(page);
+  for (size_t i = 0; i < text->size(); ++i) {
+    const base::char16& current_char = (*text)[i];
+    if (current_char == kPdfiumHyphenEOL) {
+      if (current_hyphen_position_is_valid)
+        hyphen_positions.push_back(current_hyphen_position);
+      current_hyphen_position = HyphenPosition();
+      current_hyphen_position.position = i;
+      current_hyphen_position_is_valid = true;
+    } else if (IsWhitespace(current_char)) {
+      if (current_hyphen_position_is_valid) {
+        if (current_char != L'\r' && current_char != L'\n')
+          current_hyphen_position.next_whitespace_position = i;
+        hyphen_positions.push_back(current_hyphen_position);
+        current_hyphen_position_is_valid = false;
+      }
+    }
+  }
+  if (current_hyphen_position_is_valid)
+    hyphen_positions.push_back(current_hyphen_position);
 
-  // Create a new clip path.
-  FPDF_CLIPPATH clip_path = FPDF_CreateClipPath(
-      source_clip_box.left + offset_x, source_clip_box.bottom + offset_y,
-      source_clip_box.right + offset_x, source_clip_box.top + offset_y);
+  // With all the hyphen positions, do the search and replace.
+  while (!hyphen_positions.empty()) {
+    static const base::char16 kCr[] = {L'\r', L'\0'};
+    const HyphenPosition& position = hyphen_positions.back();
+    if (position.next_whitespace_position != 0) {
+      (*text)[position.next_whitespace_position] = L'\n';
+      text->insert(position.next_whitespace_position, kCr);
+    }
+    text->erase(position.position, 1);
+    hyphen_positions.pop_back();
+  }
 
-  for (int obj_idx = 0; obj_idx < obj_count; ++obj_idx) {
-    FPDF_PAGEOBJECT page_obj = FPDFPage_GetObject(page, obj_idx);
-    FPDFPageObj_Transform(page_obj, scale_factor, 0, 0, scale_factor,
-                          offset_x, offset_y);
-    FPDFPageObj_TransformClipPath(page_obj, scale_factor, 0, 0, scale_factor,
-                                  offset_x, offset_y);
-  }
-  FPDFPage_TransformAnnots(page, scale_factor, 0, 0, scale_factor,
-                           offset_x, offset_y);
-  FPDFPage_GenerateContent(page);
+  // Adobe Reader also get rid of trailing spaces right before a CRLF.
+  static const base::char16 kSpaceCrCn[] = {L' ', L'\r', L'\n', L'\0'};
+  static const base::char16 kCrCn[] = {L'\r', L'\n', L'\0'};
+  ReplaceSubstringsAfterOffset(text, 0, kSpaceCrCn, kCrCn);
+}
 
-  // Add a extra clip path to the new pdf page here.
-  FPDFPage_InsertClipPath(page, clip_path);
+// Replace CR/LF with just LF on POSIX.
+void FormatStringForOS(base::string16* text) {
+#if defined(OS_POSIX)
+  static const base::char16 kCr[] = {L'\r', L'\0'};
+  static const base::char16 kBlank[] = {L'\0'};
+  base::ReplaceChars(*text, kCr, kBlank, text);
+#elif defined(OS_WIN)
+  // Do nothing
+#else
+  NOTIMPLEMENTED();
+#endif
+}
 
-  // Destroy the clip path.
-  FPDF_DestroyClipPath(clip_path);
-}
+}  // namespace
 
 bool InitializeSDK(void* data) {
   FPDF_InitLibrary(data);
 
 #if defined(OS_LINUX)
   // Font loading doesn't work in the renderer sandbox in Linux.
   FPDF_SetSystemFontInfo(&g_font_info);
 #endif
 
   FSDK_SetOOMHandler(&g_oom_info);
@@ skipping 1344 matching lines @@
   base::string16 result;
   for (size_t i = 0; i < selection_.size(); ++i) {
     if (i > 0 &&
         selection_[i - 1].page_index() > selection_[i].page_index()) {
       result = selection_[i].GetText() + result;
     } else {
       result.append(selection_[i].GetText());
     }
   }
 
+  FormatStringWithHyphens(&result);
+  FormatStringForOS(&result);
   return base::UTF16ToUTF8(result);
 }
 
 std::string PDFiumEngine::GetLinkAtPosition(const pp::Point& point) {
   int temp;
   PDFiumPage::LinkTarget target;
   pp::Point point_in_page(
       static_cast<int>((point.x() + position_.x()) / current_zoom_),
       static_cast<int>((point.y() + position_.y()) / current_zoom_));
   PDFiumPage::Area area = GetCharIndex(point_in_page, &temp, &temp, &target);
@@ skipping 1555 matching lines @@
     double* height) {
   FPDF_DOCUMENT doc = FPDF_LoadMemDocument(pdf_buffer, pdf_buffer_size, NULL);
   if (!doc)
     return false;
   bool success = FPDF_GetPageSizeByIndex(doc, page_number, width, height) != 0;
   FPDF_CloseDocument(doc);
   return success;
 }
 
 }  // namespace chrome_pdf